Management Research Methods 1
(MRM 1)
Pre-master UvA
Block 1
Grade 7,5
All Lecture Slides + Notes
,WEEK 1 - Data 3
WEEK 2 – Location Dispersion 12
WEEK 3 - Hypothesis Testing 28
WEEK 4 - Tests 65
WEEK 5 – Central Limit Theorem 84
WEEK 6 - Summary of Hypothesis tests and experiments 98
2
,WEEK 1 - Data
1. Data
1.1. What is data?
• Data is information
• Data has a fixed structure
o It consists of a number of properties
(variables)
§ Each column represents one
variable (case numbers)
(vertical)
o Measured from a set of
things/people/etc (units)
§ Each row represents one unit (horizontal)
The (experimental or observational) unit here is a Case. For each unit (case) we have measured several variables.
1.2. Level of measurement
How do we measure?
• Categorical (entities are divided into distinct categories): Qualitative variable
o Binary variable (two outcomes), e.g. dead or alive. Winning or losing, passing or failing 0 or 1.
§ Usually considered as nominal, but it can also be used as ordinal or discrete.
§ E.g. MALE (=yes/no) can be treated as:
• Nominal: use 2 categoreis, ignoring and denying
• Ordinal: use 2 categoreis with ordering according to masculinity
• Discrete: define MALE =1 if male and MALE =0 if female, so that MALE
counts the number of males of an individual.
o Nominal variable, e.g. whether someone is an omnivore, vegetarian or vegan. Different
types of groups, different types of things where the number doesn’t matter. Order does not matter.
§ Employed/unemployed
§ Brand of a product
o Ordinal variable, e.g. bad, intermediate, good. The order is important bad median good, low high.
§ Likert scale (1= strongly disagree, 5= strongly agree)
§ Job skill (unskilled, highly skilled)
• Numerical: Quantitative variable (or interval or scale)
o Discrete data (counts), e.g.: number of defects. Size of shoes, it is not about the category how good or
bad you are, it is the size of shoes. Other examples: throwing the dice: 1-6. The answer is fixed. Nothing
in between.
§ Number of cars sold (0,1,2,3,..)
§ Change in number of employees (..,-2,-1,0,1,2,…)
o Continuous (entities get a distinct score), e.g. temperature, body length, age, time. Can be something in
between 40,1…40,2… maximum information
§ Income (Euro, idealized view)
§ Temperature (Degrees Celsius)
If you go from continuous to binary, you lose a lot of information. If you go from binary to continuous, you receive more
and more information. He passed his exam = binary, he passed his exam with a 9,5 = continuous. So much more
information.
Hierarchy: 1. Continuous 2. Discrete 3. Ordinal 4. Nominal
Variables can be converted to a lower level of measurement. For example:
This implies a loss of information. It is not reversible.
For example: if you know that ‘body length = normal’, the exact amount of cm’s cannot be retrieved anymore.
If you transfer your data from continuous to ordinal (as in the example above) you lose a lot of information.
3
, Why is this relevant?
• For different types of data, there are different techniques to handle and analyze the data
• Over the next year. You’ll study a broad range of statistical techniques.
• The lower the amount of information in your data, the larger your sample needs to be.
Column 1: nominal; the order does not
matter. Not discrete because these are not
numbers. The order does not matter.
Column 2: nominal;
Column 3: nominal
Column 4: continuous
Column 5: continuous
Column 6: continuous
Column 7: discrete
1.3. Data collection
In quantitative research, you need to motivate and document the way you collected data.
• Is the sample representative?
• Is the data valid?
• Is there measurement error?
Population
- The complete group of interest
- All values of the relevant variables within the whole group of interest
Sample
- A (small) subset of the population for which observations are gathered
- The observed values of the relevant variables
- Random sampling: each member of the population has the same chance to enter the sample.
Exploring your data:
For qualitative data (categorical) we use:
§Frequency table
§Bar chart
§Pie chart
§Mode (most frequent outcome)
§Median, only for ordinal data (middle outcome)
For quantitative data we use:
§Histogram
§Mode, range
§Percentiles, including Median and Quartiles
§Box Plot
§Mean, Standard deviation, Skewness, Kurtosis
§Z-scores
4
(MRM 1)
Pre-master UvA
Block 1
Grade 7,5
All Lecture Slides + Notes
,WEEK 1 - Data 3
WEEK 2 – Location Dispersion 12
WEEK 3 - Hypothesis Testing 28
WEEK 4 - Tests 65
WEEK 5 – Central Limit Theorem 84
WEEK 6 - Summary of Hypothesis tests and experiments 98
2
,WEEK 1 - Data
1. Data
1.1. What is data?
• Data is information
• Data has a fixed structure
o It consists of a number of properties
(variables)
§ Each column represents one
variable (case numbers)
(vertical)
o Measured from a set of
things/people/etc (units)
§ Each row represents one unit (horizontal)
The (experimental or observational) unit here is a Case. For each unit (case) we have measured several variables.
1.2. Level of measurement
How do we measure?
• Categorical (entities are divided into distinct categories): Qualitative variable
o Binary variable (two outcomes), e.g. dead or alive. Winning or losing, passing or failing 0 or 1.
§ Usually considered as nominal, but it can also be used as ordinal or discrete.
§ E.g. MALE (=yes/no) can be treated as:
• Nominal: use 2 categoreis, ignoring and denying
• Ordinal: use 2 categoreis with ordering according to masculinity
• Discrete: define MALE =1 if male and MALE =0 if female, so that MALE
counts the number of males of an individual.
o Nominal variable, e.g. whether someone is an omnivore, vegetarian or vegan. Different
types of groups, different types of things where the number doesn’t matter. Order does not matter.
§ Employed/unemployed
§ Brand of a product
o Ordinal variable, e.g. bad, intermediate, good. The order is important bad median good, low high.
§ Likert scale (1= strongly disagree, 5= strongly agree)
§ Job skill (unskilled, highly skilled)
• Numerical: Quantitative variable (or interval or scale)
o Discrete data (counts), e.g.: number of defects. Size of shoes, it is not about the category how good or
bad you are, it is the size of shoes. Other examples: throwing the dice: 1-6. The answer is fixed. Nothing
in between.
§ Number of cars sold (0,1,2,3,..)
§ Change in number of employees (..,-2,-1,0,1,2,…)
o Continuous (entities get a distinct score), e.g. temperature, body length, age, time. Can be something in
between 40,1…40,2… maximum information
§ Income (Euro, idealized view)
§ Temperature (Degrees Celsius)
If you go from continuous to binary, you lose a lot of information. If you go from binary to continuous, you receive more
and more information. He passed his exam = binary, he passed his exam with a 9,5 = continuous. So much more
information.
Hierarchy: 1. Continuous 2. Discrete 3. Ordinal 4. Nominal
Variables can be converted to a lower level of measurement. For example:
This implies a loss of information. It is not reversible.
For example: if you know that ‘body length = normal’, the exact amount of cm’s cannot be retrieved anymore.
If you transfer your data from continuous to ordinal (as in the example above) you lose a lot of information.
3
, Why is this relevant?
• For different types of data, there are different techniques to handle and analyze the data
• Over the next year. You’ll study a broad range of statistical techniques.
• The lower the amount of information in your data, the larger your sample needs to be.
Column 1: nominal; the order does not
matter. Not discrete because these are not
numbers. The order does not matter.
Column 2: nominal;
Column 3: nominal
Column 4: continuous
Column 5: continuous
Column 6: continuous
Column 7: discrete
1.3. Data collection
In quantitative research, you need to motivate and document the way you collected data.
• Is the sample representative?
• Is the data valid?
• Is there measurement error?
Population
- The complete group of interest
- All values of the relevant variables within the whole group of interest
Sample
- A (small) subset of the population for which observations are gathered
- The observed values of the relevant variables
- Random sampling: each member of the population has the same chance to enter the sample.
Exploring your data:
For qualitative data (categorical) we use:
§Frequency table
§Bar chart
§Pie chart
§Mode (most frequent outcome)
§Median, only for ordinal data (middle outcome)
For quantitative data we use:
§Histogram
§Mode, range
§Percentiles, including Median and Quartiles
§Box Plot
§Mean, Standard deviation, Skewness, Kurtosis
§Z-scores
4
